Working type inference for all but dates/datetimes. #210.

agate/column_types.py

@@ -119,7 +119,10 @@ def cast(self, d):
         if self.date_format:
             return datetime.datetime.strptime(d, self.date_format).date()
 
-        return parse(d).date()
+        try:
+            return parse(d).date()
+        except (TypeError, ValueError):
+            raise CastError('Can not parse value "%s" to as datetime for DateColumn.' % d)
 
     def _create_column(self, table, index):
         from agate.columns import DateColumn
@@ -155,7 +158,10 @@ def cast(self, d):
         if self.datetime_format:
             return datetime.datetime.strptime(d, self.datetime_format)
 
-        return parse(d)
+        try:
+            return parse(d)
+        except (TypeError, ValueError):
+            raise CastError('Can not parse value "%s" to as datetime for DateTimeColumn.' % d)
 
     def _create_column(self, table, index):
         from agate.columns import DateTimeColumn
@@ -184,6 +190,9 @@ def cast(self, d):
 
         seconds = pytimeparse.parse(d)
 
+        if seconds is None:
+            raise CastError('Can not parse value "%s" to as timedelta for TimeDeltaColumn.' % d)
+
         return datetime.timedelta(seconds=seconds)
 
     def _create_column(self, table, index):
@@ -228,6 +237,10 @@ class TextType(ColumnType):
     """
     Column type for :class:`TextColumn`.
     """
+    @classmethod
+    def test(cls, d):
+        return True
+
     def cast(self, d):
         """
         Cast a single value to :func:`unicode` (:func:`str` in Python 3).

tests/test_inference.py

@@ -0,0 +1,79 @@
+#!/usr/bin/env python
+
+try:
+    import unittest2 as unittest
+except ImportError:
+    import unittest
+
+from agate.column_types import *
+from agate.inference import *
+
+class TestTypeInference(unittest.TestCase):
+    def setUp(self):
+        pass
+
+    def testTextType(self):
+        rows = [
+            ('a',),
+            ('b',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], TextType)
+
+    def testNumberType(self):
+        rows = [
+            ('1.7',),
+            ('200000000',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], NumberType)
+
+    def testBooleanType(self):
+        rows = [
+            ('True',),
+            ('FALSE',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], BooleanType)
+
+    def testDateType(self):
+        rows = [
+            ('5/7/84',),
+            ('2/28/1997',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], DateType)
+
+    def testDateTimeType(self):
+        rows = [
+            ('5/7/84 3:44:12',),
+            ('2/28/1997 3:12 AM',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], DateTimeType)
+
+    def testTimeDeltaType(self):
+        rows = [
+            ('1:42',),
+            ('1w 27h',),
+            ('',)
+        ]
+
+        inferred = infer_types(rows)
+
+        self.assertIsInstance(inferred[0], TimeDeltaType)